# The relationship between bias and maximum power

Discussion in 'Amp Tech Center' started by peteb, Nov 23, 2019.

1. ### petebFriend of Leo's

Posts:
4,163
Joined:
Apr 25, 2003
Location:
What limits the signal power that a tube can produce?

The tube type sets a limit. The maximum plate dissipation of a tube type some what sets an upper limit on the amount of signal power a power tube can pass. For example the max plate dissipation of a 6V6 tube is 12-14 W and the max clean power out for a pair of 6V6’s could get close but will probably never reach 24 or 28 Watts of clean power.

The bias also plays a role. A cold biased tube will probably put out less power than a hot biased tube.*

I’ve recently done some power output testing with a scope and a signal generator and what stood out to me is that you can view the signal at the speaker and you can see that there is a peak voltage that the amp will not go beyond. The wave flattens out at this invisible line, the RMS voltage can continue to rise as the wave becomes square, but the peak voltage will not rise any farther beyond this one level. The RMS voltage of a square wave is equal to the peak, so the RMS can rise up to the peak but not go above there. Thank you Elpico for showing us how this works.

What is it that sets this limit, the limit of the power of the amp? Apparently the place this clipping occurs could be the power tube or could be a pre amp tube. In a simple amp like a champ, the clipping would either have to be the power tube or the second pre amp tube stage. The first stage would never clip because it is never hit with a large signal unless a boost occurs before the 1st stage.

What actually sets the limit?

Two possibilities out of many:

The RMS voltage on the grid cannot exceed the bias voltage on the grid (or between the grid and the cathode)

The peak voltage cannot exceed the bias voltage on the grid (or between the grid and cathode)

What do you think?

I am going to do some testing on a champ power tube, and I know the testing will have one flaw. I can’t be sure if the clipping is in the power tube or a pre amp tube.

* I am not 100% sure that this is true.

Last edited: Nov 23, 2019
2. ### petebFriend of Leo's

Posts:
4,163
Joined:
Apr 25, 2003
Location:
Some math that relates to the situation, but is not the situation itself:

In theory,

A champ puts out 6 watts of clean power measured at the speaker.

4.9 V RMS is required at the 4 ohm speaker load to accomplish this

4.9*4.9/4 = 6 Watts

4.9 volts rms * 1.4 = 6.9 V peak to peak

1.4 equals the square root of 2.

A completely square wave could get up to 6.9 V rms

The max dirty power out would then be

6.9 * 6.9 / 4 = 12 W

In theory the max dirty power is twice the max clean power out.

Basically

The square root of two times a square root of two is 2, when the RMS voltage increases by a factor of 1.4, the power increases by a factor of 2.

3. ### petebFriend of Leo's

Posts:
4,163
Joined:
Apr 25, 2003
Location:
This is going to be the test,

Scope the speaker and set the amp volume to get the largest wave form prior to clipping, or the flattening out of the wave.

Read the signal voltage in RMS using a DMM at the grid and the cathode.

Assume that the clipping seen at the speaker on the scope is caused by saturation in this tube stage.

My prediction or theory or what I hope to see is

The peak to peak signal voltage on the grid minus the cathode will be equal to the DC bias voltage on the grid minus the cathode.

Or something like that.

4. ### BendyhaFriend of Leo'sSilver Supporter

Posts:
2,108
Joined:
Mar 26, 2014
Location:
Northern Germany
Spouting this false correlation again ?

The Ultimate Tone Vol. 5 – Tone Capture By Kevin O’Connor, features an amplifier, with all the circuitry and calculations explaining the how's and why's, showing that an amp can produce 30W of output with two 6V6GT tubes, or in the case of the amp, 60W with four......without the tubes running higher than the recommended maximum dissipation of 14W, in fact, they are biased safely below that level.

Then we have;
EL503 Power Pentode.
Although classified as a 27.5W tube, relating to the maximum heat dissipation, the system interrelated ratings are shown below.
This tube that was intended to be the modern replacement for the EL34, having the ability to opperate with a low voltage, and higher current.
A pair of these tubes could, when set-up right, deliver 70W of (clean as far as guitar amp go) output power, through the BIG output transformers that where used with them.

Last edited: Nov 23, 2019
5. ### petebFriend of Leo's

Posts:
4,163
Joined:
Apr 25, 2003
Location:
thanks Bendyha,

I did the test. The results were not what I expected, but not really unexpected either.

the signal on the grid when clipping first started to happen at the speaker, was not equal to the bias but about half of it.

I have the exact numbers at home but the peak of the signal was a little above half the bias voltage, and the rms signal (a better comparison to the DC bias voltage) was a little below half of the bias voltage.

This is similar to when I signal tested with a guitar. measuring the guitar generated signal at the grid of the power tubes, it is typical that the guitar signal will reach as a maximum about half of the bias voltage. and in extreme cases the guitar signal can equal the bias voltage, and in extreme extreme conditions the guitar voltage will slightly exceed the bias voltage. your mileage may vary

6. ### elpicoTele-Holic

Posts:
777
Joined:
Sep 14, 2011
Location:
Vancouver BC
If we're talking about a standard push-pull power amp then it's the other way around. Colder bias = more potential power output. But it doesn't happen automatically.

If the question is "what determines the max clean power" in that type of amp then the answer is "several things":

- Tube type (and number of pairs)
- Supply voltage
- Bias

It's an interactive relationship between those four.

There are a few other gotchas, things there "needs to be enough of" like the amount of steel in the OT, it's primary inductance, it's ability to carry current, the ability for the tubes to radiate heat away into the environment etc. If you're buying an OT rated for the power you're trying to achieve and you don't pack the tubes in an insulated box then you can mostly forget about these things.

Last edited: Nov 26, 2019
LudwigvonBirk, Modman68 and Bendyha like this.
7. ### petebFriend of Leo's

Posts:
4,163
Joined:
Apr 25, 2003
Location:
Thanks Elpico,

I agree, a colder bias will create the potential for more power out, and I agree that it doesn’t happen automatically, and I would say that just cooling the bias will not increase the power, it might lower it. Only when the signal is raised to match the raised bias voltage, will the power increase. And the power increases because the signal increased.

Why would push pull versus single ended matter?

I agree with this but I’m going to add one more thing.

The plate voltage that powers the tube has to be up to spec for the tube type.
The load has to be right.
The negative bias voltage needs to be as large as it is supposed to be for the tube type and plate voltage and load.
The signal needs to match up to the bias level.

I’m pretty certain the ratio of signal level to bias level, or the percentage of the bias voltage that the signal reaches, is one of the key parameters that describe tube operation. It is a useful parameter when determining when the tube will pass signal and when it won’t, and what the max power out available is.

Last edited: Nov 26, 2019
8. ### elpicoTele-Holic

Posts:
777
Joined:
Sep 14, 2011
Location:
Vancouver BC
Because single ended amps are class A

LudwigvonBirk likes this.
9. ### petebFriend of Leo's

Posts:
4,163
Joined:
Apr 25, 2003
Location:
I meant to show voltages, here they are.

This is an AA764 champ amp.

Bias voltage: 22.8 VDC

Grid signal at clip: 9.25 VAC RMS
Cathode signal at clip: 0.4 VAC RMS
Grid minus cathode: 8.85 VAC RMS

Comparing grid signal at clip in VAC RMS to bias voltage in DC:

8.85/22.8 = 39%

Comparing grid signal at clip in peak voltage to bias voltage:

12.5 / 22.8 = 55%

it appears that clipping begins when the signal on the grid is approximately half the the bias voltage.

say more, I'm always interested in the differences between single ended versus push pull, or class A versus class AB

10. ### elpicoTele-Holic

Posts:
777
Joined:
Sep 14, 2011
Location:
Vancouver BC
There must quite a few threads covering that topic already. The short story is push pull amps can be designed to run class A like a single ended amp, but they can also do something single ended amps can't which is split the job in half and allow one tube to "rest" while the other is having a turn. Taken to an extreme in class B, the tubes work at double their rated power during their turn because they get to cool down during their rest.o

To achieve this trick the tubes are biased so cold at idle they're esentially off, the opposite of class A where they idle at max power. That's why I said in a push pull amp colder bias has the potential to make more power. Single ended amps can't do this trick so it's not true there.

LudwigvonBirk likes this.
11. ### petebFriend of Leo's

Posts:
4,163
Joined:
Apr 25, 2003
Location:
Thanks Elpico,

I see what you are saying, a colder bias makes it so the tube works less of the time so that it can work harder when it is working, going from class AB operation toward class B operation does this.

Another way of looking at the same thing is, that when the operating point is lowered to cutoff, it provides more room to raise the signal voltage and still stay below saturation or the bias volatage.

That is a big advantage class AB has over class A. Class AB only has to fit one half of the wave, or a little bit more than one half, in between cutoff and saturation, where class A has to fit the upper half and lower half in between cutoff and saturation, making the class AB signal almost double the class A signal.

12. ### LudwigvonBirkTele-Holic

Age:
117
Posts:
865
Joined:
Aug 26, 2017
Location:
Elpico didn’t actually say all this.

13. ### petebFriend of Leo's

Posts:
4,163
Joined:
Apr 25, 2003
Location:
And, I see further evidence of what Elpico said, lower bias for more clean power out. The champ with the signal shown in the images above, is biased hot and struggles to show 5 or 6 clean Watts on the scope. Where as my Princeton is biased very cold but seems to have no problem showing 12 clean Watts on the scope.

14. ### petebFriend of Leo's

Posts:
4,163
Joined:
Apr 25, 2003
Location:
Full signal before clipping

Saturation

Cutoff

All of them

Last edited: Nov 27, 2019
15. ### kboldTele-Holic

Posts:
811
Joined:
Jul 21, 2015
Location:
Australia
In a nutshell: plate voltage.
It's ohms law. Increase the voltage and you increase the power.
Consider the old Dynacord Eminent II amp, with a plate voltage over 750V, outputting 80 Watts from 2 EL34's.

16. ### petebFriend of Leo's

Posts:
4,163
Joined:
Apr 25, 2003
Location:

I agree, it starts with the plate voltage.

Class B?

I would like the know how that amp is biased, and the signal level supplied to the power tubes.

I have heard about class B amps and how the output Wattage is big and the plate voltage too, but I have never heard of or learned of the bias or operating point.

17. ### kboldTele-Holic

Posts:
811
Joined:
Jul 21, 2015
Location:
Australia
So, according to the schematic, adjustments are:
Anode (plate) = 740V Cathode =0.3V (= 30mA) (IIRC I measured 770V at plate)
Bias is adjusted for -36V at input. Actual signal level at EL34 input = 27V (i.e. signal as opposed to DC adjustment)

At these values, it's biased 'hot', but I think the 80 Watts (I googled) is closer to 50 Watts.
IIRC I biased at around 22 to 25mA. Aahhh - cool running for a Dynacord.

Still, more Volts = more Watts still stands true.
I should also note that, for this amp (Dynacord), most new EL34's made today won't stand up to this punishment (NOS tubes seem to do OK though).
I don't think they're made today to actually run at or near 800V.

18. ### petebFriend of Leo's

Posts:
4,163
Joined:
Apr 25, 2003
Location:
Thanks kbold, the dynacord is certainly worth checking out.

750 volts is high but OK

80 Watts from 2 EL34s is high

19. ### petebFriend of Leo's

Posts:
4,163
Joined:
Apr 25, 2003
Location:

Thank you Kbold,

I have been interested in knowing the operating point of these high powered/low tube count amps.

These are my observations.

In order of importance,

750 volts on the plate is really high

Bias of -36 V is probably OK for class AB

Signal of 27 volts is good compared to -36 volt bias.

Current 30 mA. Fine. Definitely not class B which is idle biased at 0% of max plate dissipation

Plate dissipation is 750 V x 0.03 A = 22.5 W which is 125% of an EL34 with max plate dissipation of 20 W.

125% of MPD indicates the tubes are in or close to red plate condition.

A max power out of 80 Watts when two EL34s max dissipate only 40 W could only maybe make sense in class B but not in class AB. The only thing the amp has going for it is really high plate voltages and it really should be red plating.

20. ### kboldTele-Holic

Posts:
811
Joined:
Jul 21, 2015
Location:
Australia
Yes, that's why I don't believe the "Google " claimed 80 Watts (I believe someone misread 50W as 80W)
And yes, 50W is still a lot of Watts for 2 x EL34's.
Those Germans are really clever. Maybe the whopping transformer helps? Maybe the horizontally mounted EL34's help thermal dissipation? I never got to turn the volume pot anywhere near these ratings - I have neighbors.
On refurb' I did replace the EL34's: one of the 2 had suffered red plating. This pair still work, and are now happily in retirement in my current amp, cruising at a comfy 500V / 30mA.

IMPORTANT: Treat everyone here with respect, no matter how difficult!
No sex, drug, political, religion or hate discussion permitted here.